Wind Power From Kites

Transcription

Wind Power From Kites
Project #: DJO - 0107
WIND POWER FROM KITES
A Major Qualifying Project Report
Submitted to the Faculty
of the
WORCESTER POLYTECHNIC INSTITUTE
in partial fulfillment of the requirements for the
Degree of Bachelor of Science
In Aerospace Engineering
SUBMITTED BY:
Michael R. Blouin Jr.
[email protected]
Benjamin E. Isabella
[email protected]
Joshua E. Rodden
[email protected]
Date: April 27th, 2007
Professor David Olinger, Project Advisor
Abstract
The goal of this project was to study the feasibility of using tethered kites to
generate power from the wind. Generating electricity using kites instead of wind turbines
may have certain advantages, particularly for developing nations. These include
generating power at low cost while eliminating certain environmental problems
associated with wind turbines. Another advantage is that kites can fly at greater heights
than wind turbines can operate. Since wind speed increases with height, and available
power is proportional to wind speed cubed, the wind power potential is larger for kites. A
literature review was conducted of previous studies of kite power systems. A mechanism
was designed to convert the oscillating tether tension caused by the vertical motion of a
kite into rotary shaft motion to drive a generator. The mechanism is based on a rocking,
balanced beam of 5 meter length attached to a power train and Sprag clutch. The design
uses a commercially available sport kite which is 10 meters squared in size. A previously
developed MATLAB code was used to model the final system design, and power outputs
comparable to small wind turbines were predicted.
2
Acknowledgements
The Wind Power from Kites group would like to thank:
Professor David J. Olinger
For his guidance, leadership, encouragement, and enthusiasm on this project from the
beginning of development.
Dr. J. S. Goela
For originally conceiving the concept of wind power from kites as well as acting as
technical consultant throughout the duration of the development of this project.
Andrew Ghezzi, owner of Powerline Sports in Seabrook, NH
For his insight into kite dynamics and the sport of kite boarding as well as the instruction
for proper operation of our kite.
Professor Gatsonis, Professor Blandino, Professor Demetriou, and Professor Hussein
For constructive criticism and guidance presented during weekly AE MQP presentations.
Fellow Aerospace Engineering Students
For constructive criticism and encouragement during weekly AE MQP presentations.
Dale Perkins
For allowing the use of Overlook Farms as the primary test site for the kite power
demonstrator.
The Staff of the WPI Machine Shops
For assistance in dealing with minor issues with part tolerances in a timely fashion.
Col. Ken Stafford
For introducing us to the idea of using a sprag clutch in the final mechanism design.
3
Table of Contents
Abstract ......................................................................................... 2
Acknowledgements ....................................................................... 3
Table of Figures ............................................................................ 5
List of Tables................................................................................. 7
1. Introduction .............................................................................. 8
2. Background ............................................................................. 16
2.1 Previous Studies................................................................................. 16
2.2 Potential Wind Power ....................................................................... 20
3. Methodology............................................................................ 23
3.1 Kites .................................................................................................... 23
3.1.1 Kite Purchase ................................................................................................... 26
3.1.2 Kite Testing...................................................................................................... 28
3.2 System Design .................................................................................... 30
3.2.1 Possible Mechanism Designs........................................................................... 30
3.2.2 Mechanisms Evaluation ................................................................................... 39
3.2.3 System Design ................................................................................................. 48
3.2.4 System Construction ........................................................................................ 53
3.3 System Simulations............................................................................ 61
3.3.1 Steady State Simulations.................................................................................. 62
3.3.2 Dynamic Modeling Simulation........................................................................ 68
3.4 System Stress Analysis ...................................................................... 74
4. Results...................................................................................... 78
4.1 Kite Testing ........................................................................................ 78
4.2 Determination of Simulation Values ............................................... 84
4.2.1 Steady Sate Simulation Results ....................................................................... 84
4.2.2 Dynamic Simulation Results............................................................................ 89
4.3 System Stress Analysis ............................................................... 98
5. Conclusions............................................................................ 102
5.1 Future Recommendations .................................................. 104
References: ................................................................................ 107
APPENDIX A: Drachen Foundation ...................................... 108
APPENDIX B: Heifer International’s Overlook Farm......... 109
4
Table of Figures
Figure 1 - Power output and wind velocity for turbine or kite with A = 10 m2 area. ...... 10
Figure 2 - KiteGen concept............................................................................................... 11
Figure 3 - David Lang's Reel Concept.............................................................................. 12
Figure 4 - SkySail ............................................................................................................. 14
Figure 5 - Early Goela Kite Models, From Goela (1983)................................................. 18
Figure 6 - Goela Kite Model, From Goela (1983) ............................................................ 18
Figure 7 - Goela Spring Mechanism View 1, From Goela (1983) ................................... 19
Figure 8 - Goela Spring Model View 2, From Goela (1983)............................................ 19
Figure 9 - Wind Power based on Wind speed................................................................... 20
Figure 10 - New England Wind Speed 30m above Ground ............................................. 21
Figure 11 - New England Wind Speed 100m above Ground ........................................... 21
Figure 12 - Peter Lynn Venom and Twinskin Technology explained.............................. 26
Figure 13 - Peter Lynn Guerilla ........................................................................................ 27
Figure 14 - Beach Kite Testing......................................................................................... 28
Figure 15 - Angle of Attack Approximation..................................................................... 29
Figure 16 - Modern Kite Control System in a De-Powered State .................................... 31
Figure 17 - Control Mechanism in a "Powered" State...................................................... 32
Figure 18 - Simple Lever .................................................................................................. 33
Figure 19 - Simple Lever Reverse Clutch......................................................................... 34
Figure 20 - Possible Two-Kite System #1 ........................................................................ 35
Figure 21 - Possible Two kite System #2 ......................................................................... 35
Figure 22 - Oil Pump Jack ................................................................................................ 36
Figure 23 - Example Sprag Clutch.................................................................................... 37
Figure 24 - Pumpjack/Sprag Clutch Combo..................................................................... 38
Figure 25 - Possible Roll Control Spring Mechanism ...................................................... 39
Figure 26 - Illustration of Pump Jack Design ................................................................... 40
Figure 27 - Illustration of Sprag Clutch Design................................................................ 41
Figure 28 - Illustration of Sprag Clutch / Pump Jack Combo Design .............................. 41
Figure 29 - AutoCAD system Model................................................................................ 49
Figure 30 - Telescoping Beam.......................................................................................... 50
Figure 31 - Exploded View Angle of Attack Mechanism ................................................ 50
Figure 32 - Exploded View Gear System ......................................................................... 52
Figure 33 - Top Side of Structure ..................................................................................... 54
Figure 34 - Structure Braces and Beam Fulcrum Point .................................................... 55
Figure 35 - Frontal View of Structure............................................................................... 55
Figure 36 - Turnbuckle Assembly .................................................................................... 56
Figure 37 - Balanced Arm Assembly................................................................................ 57
Figure 38 - Beam fulcrum point assembly........................................................................ 58
Figure 39 - Angle of Attack Change Mechanism Assembly ............................................ 60
Figure 40 - Axles and Gears Assembly ............................................................................ 61
Figure 41 - Kite Model, From Goela (1983)..................................................................... 62
Figure 42 - Pivoting Arm Diagram................................................................................... 66
Figure 43 - Two Gear-System Schematic......................................................................... 67
5
Figure 44 - Modified System Simulation.......................................................................... 69
Figure 45 - Simulation Parameters ................................................................................... 70
Figure 46 - Simulation Parameters Part #2 ....................................................................... 72
Figure 47 - Rotating Beam reaches a set Angle in the program ....................................... 74
Figure 48 - Beam Stress.................................................................................................... 75
Figure 49 - Top Support Beam ......................................................................................... 76
Figure 50 - Beam Stress Diagram..................................................................................... 76
Figure 51 - Ozone Access 4m........................................................................................... 78
Figure 52 - Cabrinha Crossbow 7m .................................................................................. 80
Figure 53 - Spring Deflection vs. Weight (N) .................................................................. 83
Figure 54 - Kite Line Tension vs. Lift over Drag............................................................. 85
Figure 55 - Pivoting Beam Power vs. Angle of Attack .................................................... 86
Figure 56 - Chain Tension vs. Lift over Drag................................................................... 87
Figure 57 - Torque vs. Lift over Drag............................................................................... 88
Figure 58 - Average Coefficient of Power vs. Counterweight (N)................................... 89
Figure 59 - Counterweight of highest Power Production ................................................. 90
Figure 60 - Highest Coefficients of Power for Wind Speed............................................. 91
Figure 61 - Optimal Power Production for given wind speed .......................................... 92
Figure 62 - Yearly power production vs. average wind speed ......................................... 93
Figure 63 - Kite Motion .................................................................................................... 94
Figure 64 - Simulated Angle gamma (deg.) vs. Time (s) ................................................. 95
Figure 65 - Simulated Line Tension (N) vs. Time (s)....................................................... 96
Figure 66 - Simulated Lift over Drag vs. Time (s) ........................................................... 97
Figure 67 - Simulated Cl vs. Cd........................................................................................ 98
Figure 68 - Stress Vs. AOA .............................................................................................. 99
Figure 69 - Deflection Vs. AOA..................................................................................... 100
Figure 70 - Support Beam vs. Kite Angle of Attack....................................................... 101
Figure 71 - Beam Deflection vs. Angle of Attack .......................................................... 101
6
List of Tables
Table 1 - Peter Lynn Guerilla 10m^2 characteristics ....................................................... 27
Table 2 - Evaluation Scale ................................................................................................ 42
Table 3 - Evaluation Matrix.............................................................................................. 45
Table 4 - Ozone Axis Kit Characteristics ......................................................................... 79
Table 5 - Ozone Axis Tension Data.................................................................................. 79
Table 6 - Cabrihna Cross Kite Characteristics.................................................................. 81
Table 7 - Collected Data for Cabrihna Crossbow 7 m^2.................................................. 81
7
1. Introduction
There are several sources of renewable energy currently in use today including
bio-fuels, solar energy, hydrogen fuel cells, and wind power. The focus of this project
will be on wind power and its potential use in a developing nation. Wind power
production provides several significant benefits when being considered for integration
into the infrastructure of a developing nation or a rural area. Wind turbines are generally
used to generate power from the wind. Wind power systems, such as wind turbines, do
not need long transmission lines and therefore can create and distribute electrical power
outside the electrical grid.
Though it is apparent that there are significant advantages to employing wind
power systems in developing nations, there are a number of significant disadvantages to
the current primary source of wind energy, the wind turbine. When considering use in a
developing nation, a wind turbine is a very expensive commodity. Wind turbines include
several components; a set of fan blades, a large tower, a power conversion system, and a
battery storage unit. Another significant drawback for the wind turbine is the impact that
it has on the environment. Even though the wind turbine eliminates the greenhouse
emissions typical of fossil fuel systems, wind turbines produce a considerable amount of
noise pollution, visual pollution, and are believed to be a significant cause of bird deaths.
The concept of a kite powered wind energy system can eliminate these
disadvantages while retaining the advantages of a wind power system. This proposed
system would convert wind power into a usable form of energy, either electrical or
mechanical. For the production of electrical energy, the kite system will employ a system
to convert the up and down linear motion of a kite tether, as the angle of attack is
8
changed, into a rotary motion of a shaft. Connection of the shaft to a generator will
produce an electrical current that can be stored in batteries.
A kite driven energy system has the potential to have significantly lower costs
than a wind turbine as there is no need for the large tower that a wind turbine requires.
Without the large turning blades of the wind turbine the noise pollution and bird hazard
concerns are eliminated. The concern with visual pollution is also eliminated as kite
systems would not negatively impact attractive natural landscapes.
The kite powered system can be implemented in areas with larger variations in
wind velocity as the kite has the ability to reach much greater altitudes. The advantage of
reaching much greater altitudes is that at higher altitudes the wind velocity is higher and
more constant with less gusting. A generally recognized 'rule of thumb' is that wind speed
increases to the 1/7th power of the change in height above ground. Figure 1 is a
comparison between the potential power output and the wind velocities available to a
10m^2 kite and wind turbine. Power increases proportionally with the cube of the wind
velocity, P α V3, and therefore increases in altitude result in greater increases in power
output potential.
9
1400
20
1200
Power (Watts)
800
10
600
400
Wind Velocity (m/s)
15
1000
5
200
0
Kites
Turbines
0
0
20
40
60
80
100
120
140
160
Altitude (meters)
Figure 1 - Power output and wind velocity for turbine or kite with A = 10 m2 area.
1.1 Previous Research
Fossil fuel exhaustion concerns during the past few decades have given rise to
renewed interest in renewable energy sources. Therefore, there has been an increasing
amount of research done into different types of renewable energy sources. Wind power is
believed to be a beneficial form of renewable energy. With the goal of decreasing the
costs of this already cheap form of energy, scientists and enthusiasts alike have been
studying the feasibility of using kites to harness wind power and convert it into a useful
form of mechanical or electrical energy.
There have been a large number of concepts and designs created for systems
intended to convert wind power into electrical energy as well. At this time it seems as
though the majority of systems that have been designed for the production of electrical
energy have been extremely complex and large in size. One example of a very large kite
power system is the “KiteGen”, as seen in Figure 2. It is believed that this system
designed in Italy has the potential of replacing a nuclear power plant. The KiteGen
10
resembles a very large merry-go-round with stacked kites attached to the outer rim so as
to turn a large platform, producing the electrical energy. (sequoiaonline.com, 2006)
(http://www.wired.com/science/discoveries/news/2006/10/71908)
Figure 2 - KiteGen concept
Lang (2005) compares and studies the feasibility of several different concepts for
systems designed to convert wind power into electrical energy using kites. Lang uses a
detailed decision matrix to compare six different concepts for system designs and
ultimately determined what he objectively believed to be the most feasible options. Of the
six concepts compared, the two top scoring concepts were the “KiwiGen”, and the
“Reel.” The KiwiGen is the same concept as the KiteGen described above. The Reel,
shown in Figure 3, is a concept utilizing a series of mechanical and electrical components
to harvest energy from the kite as it is pulled out and wound back in. The kite in the Reel
concept is pulled out when it is at a high angle of attack until the angle of attack is
mechanically changed and the kite is pulled in by another mechanical component. The
11
resultant energy is the energy gained from the turning gears as the kite is pulled out
minus the mechanical energy required to reel the kite back in to restart the cycle.
Figure 3 - David Lang's Reel Concept
Above are examples of relatively recent developments in wind power harnessing
to produce a usable form of energy, but there were numerous studies done in the 1970’s
and 1980’s as well. Loyd (1979) analyzed the power potential of large area kites in
crosswind conditions. This study developed equations of kite motion when interacting in
the crosswind direction with the understanding that potential power output is maximized
when the kite angle of attack is at a ninety degree angle to the direction of the wind. Loyd
estimated that a kite with an area of 2000 m^2 and 1200 m tethers has the potential of
producing an average power output of 45 MW.
Dr. J. S. Goela, of the Indian University of Technology Kanpur, investigated the
feasibility of a kite system to convert wind power into a usable form of mechanical
energy as well as electrical energy. Goela (1983) is one of several yearly reports
12
chronicling the progress of Goela’s research. The work in this particular publication
focuses on harnessing wind power to lift a bucket of water from a deep well. After much
work in the area of kite wind power, Goela proposed to design and test a system to
confirm the ability of a kite system to translate wind power into usable energy. Goela
(1983) stated that the dynamical analysis of kite motions and the design of the simple
mechanical system proved that the conversion of wind power into either mechanical or
electrical energy is a feasible option for renewable energy production. Goela’s work will
be further summarized in later sections.
When one considers the sport of kiteboarding, it is apparent that wind power is
already being converted into a usable form of mechanical energy. Kiteboarding is a sport
in which a kite is used to pull a rider on land, sea, or water. A system called a “skysail” is
another kite system intended to translate wind power into a useful form of mechanical
power. A skysail, shown in Figure 4, is a large kite on the order of 5000 m^2 that is
believed to have the ability of decreasing the required fuel of an ocean-going tanker by as
much as 20%, saving billions of dollars on diesel fuel purchases. (foxxaero, 2003)
13
(http://www.diseno-art.com/encyclopedia/archive/skysail.html)
Figure 4 - SkySail
1.2 Project Goals
The goal of this project was to design and construct a wind power from kites
demonstrator. This demonstrator would have to consist of several key components with
particular functions. The demonstrator would have to have the ability to change the angle
of attack of the kite as well as convert the linear motion of the kite into a rotary motion
via a power conversion mechanism.
In order for the kite to be used to convert wind power into electrical power there
must be a minimum of three mechanisms incorporated into the particular system. These
mechanisms are: main mechanism to change linear motion into rotary motion; angle of
attack change mechanism; and kite stability control mechanism.
The main mechanism consists of several components to be discussed later
operating in unison to convert the linear vertical motion of the tether into a rotary motion
14
of a shaft. The angle of attack changing mechanism is intended to work with the main
mechanism design to change the angle of attack of the kite in order to force it to move up
and down in the air, thus creating the linear force that the mechanism converts to a rotary
motion. The concept of a stability control mechanism arose from the belief that a
mechanism may be required to control and limit the adverse effects the wind may have
on the kite due to unexpected direction changes and gusts. Though the issue of stability
was addressed, the mechanism was not completely designed nor was it incorporated into
the final system design. Each of the components of the system will have to work in
concert in order to efficiently convert the linear motion of the kite into a rotary motion to
produce energy. The design and construction of each individual subsystem will be
described in detail in later sections.
In order to analyze the potential power output of the designed system several
forms of simulations were performed. As previously stated, the steady state equations of
kite dynamics were developed from the work of Dr. Goela. The steady state equations
were used to simulate the kite in a stationary position to determine potential line tensions
and power output. Dynamic simulations were also done in order to couple the steady state
simulations with the specific geometries and dimensions of the designed system. To
ensure that the forces applied to the designed system were not in excess a series of stress
analysis calculations were performed to determine the strength of the system at several
critical points. Simulations will be discussed in much more detail in future sections. In
the design process software programs such as SolidWorks and Working Model were used
in order to model the individual mechanisms and the system as a whole as the design
progressed.
15
2. Background
2.1 Previous Studies
Research done in the past on the concept of using a kite for power generation was
primarily done by Dr. Goela. Therefore, much of the background information providing
the theory for this project has come from one of three publications by Dr. Goela; Goela
(1983), Goela (1979), Goela et al. (1986). The publication that has provided much of the
analytical background and the equations of motion for dynamic simulations, which will
be introduced later in this report, is Goela (1983). This particular publication was one of
several yearly reports compiled by Dr. Goela and his research assistants at the Indian
Institute of Technology Kanpur. Dr. Goela has served as a technical consultant on this
MQP project.
Goela (1983) performs analysis of the steady state motion of the kite during both
stages of its motion, ascent and descent. The mathematical analysis concentrated on the
forces acting on the kite and the forces produced by the kite on the system as a result of
the kite motion. With the equations of kite dynamics formulated, Goela (1983) considers
several other important factors in the analysis of the kite system. Goela (1983) determines
the relative efficiencies of the kite system, such as the potential power coefficients, as
well as the delay time between the phases of ascent and descent. .
Goela (1983) also studied the design of the kite and the mechanism to be used for
the system to convert wind energy into mechanical energy. In order to develop the best
kite design for their purposes, Goela and his team tested several different designs for the
kites. In order to collect data, the kites were tested in a large wind tunnel and specific
force measurements were taken at different angles of attack. Several different types of
kites (Figure 5) were tested in order to find the one that best suited the objective of the
project and the final choice was a conyne kite (Figure 6). This type of kite gave the best
overall results for the desired properties as “it incorporates the lifting advantage of a flat
kite with the stability of a box kite” Goela (1983).
The second main feature of a kite-powered system is the mechanism that
translates the motion of the kite into a usable form of energy. In the case of Goela (1983),
the form of usable energy is the ability to raise water out of a well. The mechanism that
Dr. Goela and his team designed consisted of a balanced beam on a fulcrum with springloaded assists as shown in Figure 7 and Figure 8. The springs in the system were used as
a switching mechanism in order to chance the angle of attack of the kite, changing from
ascent to descent. As the balanced beam reaches the top of its path the water is discarded
from the bucket, decreasing the weight of the bucket as the angle of attack is decreased
with the flip of the lever. The motion described above is portrayed in the two stage view
in Figure 8. Once the angle of attack is changed the bucket is slightly heavier than the
tension in the tether and the kite is pulled back down to its starting point. The cycle
restarts once the lever is triggered in the opposite direction during the descent of the
bucket and kite. The system of Goela (1983) was intended to lift a bucket full of water
from a well and therefore cannot be directly incorporated into the work being done for
this project, although as we shall see, it did influence our demonstrative analysis.
17
Figure 5 - Early Goela Kite Models, From Goela (1983)
Figure 6 - Goela Kite Model, From Goela (1983)
18
Figure 7 - Goela Spring Mechanism View 1, From Goela (1983)
Figure 8 - Goela Spring Model View 2, From Goela (1983)
19
2.2 Potential Wind Power
When gauging the potential power of wind systems it is important to remember
how this power is calculated. The equation for wind power is calculated using:
(1)
The Vwind is the velocity of the wind, ρ is the density of the air and A is the area
swept by the wind. From this equation it is evident that an increase in wind velocity
results in a much higher wind power. Using the wind power equation, with potential wind
power based on sea level conditions (ρ = 1.23 kg/m^3), and an area (A = 1 m^2):
Figure 9 - Wind Power based on Wind speed
With this in mind, it becomes clear that wind power systems generate much more
power in areas of high wind speed. This is why many wind turbines reach as high as 50m
in height. The speed of wind tends to improve upon at higher altitudes. For specific areas,
wind charts have been generated up to 100m in height. These wind charts are based upon
20
numerical data that has been collected and average. Here is an example of a wind chart
for the New England area detailing average wind speeds at a height of 30m.
Figure 10 - New England Wind Speed 30m above Ground
Similarly, here is a chart of average wind speeds for the New England area at a
height of 100m.
Figure 11 - New England Wind Speed 100m above Ground
21
Comparison of the two heights shows that the average wind speed increases by 11.5 m/s in a change of height of 70m. For heights greater then 100m not much data is
available. For these higher heights it is generally acceptable to use the equation of the
form:
(2)
Of course, most wind turbines cannot reach heights greater then 50m, where the
potential power of the wind is much higher. There is where the use of kites can unlock
the potential power found at high elevations. Kites can fly at heights up to thousands of
feet depending on the area. In the United states the maximum height is regulated to 500
feet (152 meters) based on Federal Regulations. Most third world countries have little to
no restrictions on the possible height of the kite. However, the potential power output at
even 500 feet is still much higher than 150 feet (50m).
An example demonstrates the potential power at higher elevations. Consider two
separate elevations with one at 30m, and another representing 100m. Using wind data
from earlier graphs, an area of Massachusetts experiences 4.5m/s average wind speeds at
30m elevation. While the 100m elevation experiences an average wind 6.5m/s.
Calculating the wind power per one square meter gives shows that 56 W/m^2 is
potentially available at 30m, while 156 W/m^2 is available at a height of 100m. This
change in height represents a 300% increase in the potential power. The design of our
project is to harness this extra available power.
22
3. Methodology
The main goal of this project is to design a wind power from kites demonstrator
that can be proven to generate useful amounts of electricity in a costly and efficient
manner. To accomplish this goal our project was split into two design phases. First is the
choice of a suitable kite that can be easily manipulated and that is readily available. The
second is a mechanism that can successfully translate the linear oscillating motion of a
kite into a rotary motion that can power a generator.
3.1 Kites
For our applications, the properties we are looking for in a kite are: stability, ease
of use, and durability. Our original idea was to fabricate our own kite so we could design
it to exactly fit our needs but due to the time involved we rejected this idea in favor of
using a tested and proven kite design. We chose to use a kite that was designed
specifically for kite boarding, a sport where the rider is towed across the ground or water
by a kite, relying only on a light wind for acceleration. These kites come in a variety of
sizes, ranging from 2 square meter trainer kites to 20 square meters for heavier riders,
with most riders using between a kite between 9 and 13 square meters. The kites operate
in what is called the power zone, or the area in which the kite gets the most lift. The
power zone is typically defined as a half-hemisphere stretching 180 degrees in front of
the rider and projecting 90 degrees down in an arc from the vertical direction. The
advantage of a large kite is that it will provide line tension at higher wind speeds and is
more stable when aloft due to its greater area. The smaller kites have an equal advantage
23
at lower speed winds, requiring only a light breeze to keep them aloft, but are less stable
due to their smaller profile.
Kite boarding kites employ several different control systems, ranging from the
simple 1 line system of a traditional kite to a 5-line system of some of the more advanced
kite-boarding kites. Typically, most kite boarding kites have either a 2 or 4 line
configuration. In a 2-line setup, two lines are attached to the kite, usually to a bridal
system that then spreads out to the underside of the kite. The downside to this system is
that only the roll motion of the kite can be altered, not the angle of attack. Because of
this, we immediately rejected any kite with this type of system, since our project relies on
changing the kite’s angle of attack. The other line configuration, a 4-line system, has a
line at each corner of the kite.
The most commonly used type of kite boarding kite is the traditional parafoil.
This type of kite consists merely of two sheets of rip-stop nylon with vertical strips of
nylon uniting the two. The air is allowed to move between the top and bottom layer of
the parafoil and gives it its structure. This kite is normally found in a 2-line variety, but
can come in 4-line. In kite boarding, this type of kite is traditionally used as a trainer kite
to train riders how to control the kite. The downside of this kite was that it was unstable,
requiring constant supervision and roll adjustments to keep it within the power zone.
A Lead Edge Inflatable, or LEI kite, was also researched. The LEI features a
leading edge and supporting struts that can be inflated and pressurized to provide rigidity
and create an airfoil-like shape to the kite. However, LEI kites have only rip-stop nylon
sheets in between the inflated struts, which cause increased drag due to the ability of the
kite to flex, changing the geometry of that trailing edge and causing it to flutter up and
24
down. The advantage of this particular kite for kite boarding is that if the kite hits the
water, it will stay above the surface due to the ballast from the inflatable edges. We
considered this type of kite with the intention of filling the leading edge and struts with
helium in order to keep the kite aloft in the event that a constant wind was lost. We later
rejected this idea due to the amount of helium needed to keep the kite aloft, and exotic
materials that would have to have been implemented to ensure the helium did not leak out.
Overall, these kites were generally stable, but did require constant trim to keep them in
what is called the power zone.
Another kite commonly used by kite boarders was a Bow kite, which is a slightly
modified version of the LEI. With a bow kite, the leading edge lines run across the
interior body of the kite, giving the rider the ability to control the front profile of the kite,
thus reducing or increasing the effective surface area that the wind can act upon and
thereby increasing the kites operating wind ranges. These kites however are notoriously
unstable and sensitive, and as a result, we eliminated them as a choice due to the amount
of attention they needed to be operated.
Kites manufactured by Peter Lynn, Inc. were also considered. The design of this
kite utilizes an unpressurized inflatable structure, whose profile mimics that of an airfoil.
The advantage of this design over a lead edge inflatable kite is that the since the kite
mimics an airfoil, it has an uninterrupted air flow pattern over both sides of the kite, as
opposed to a LEI kite, which has a continuous line over the top edge, but a stagnation
point on the bottom edge shortly after the air passes the front bladder, as seen in Figure
12.
25
Figure 12 - Peter Lynn Venom and Twinskin Technology explained
Additionally the Venom also has a feature called an auto zenith, which enables
the kite to re-launch itself in the event that it falls out of the power zone. This feature
also ensures that when the kite is at rest and without any rider intervention, the kite stays
at the zenith with minimal perturbations.
3.1.1 Kite Purchase
As mentioned in the previous section, a four-line kite provides enough dynamics
to keep the kite stable and under control. Four lines are also necessary to ensure that the
kite’s angle of attack can be successfully controlled. More than four lines would be
unnecessary because the addition of more lines provides little to gain in the kite’s control.
With four lines and an auto-zenith stability feature, it was found that a Peter Lynn Twinskin kite would be best for the purposes of the project.
26
Peter Lynn Kites are very difficult to find at retailers and very expensive for the
new factory versions. To reduce costs it was decided to find a relatively cheap, rarely
used Peter Lynn kite from either the kite boarding community or a host of various
Internet locations. Eventually, a Peter Lynn Guerilla 10 m^2 kite was found and
purchased on EBay for $300.00. According to the seller the kite was flown only once,
and personal inspections of the kite showed that it had very little use. This kite was
purchased for a relatively low price as a new Peter Lynn Guerilla costs approximately
$800.00. Figure 6 shows the 12m^2 version of the Peter Lynn Guerilla.
Figure 13 - Peter Lynn Guerilla
Some of the Kite Specifications:
10 m^2
Kite Area (m^2)
22.24
Kite Weight (N)
1.575
Chord Length (m)
7.62
Wingspan (m)
4.84
Aspect Ratio
Four
# of Control Lines
Minimum Required Wind Speed (m/s) 3 m/s
13 m/s
Maximum Wind Speed (m/s)
Table 1 - Peter Lynn Guerilla 10m^2 characteristics
27
3.1.2 Kite Testing
To become more familiar with the use and control of the kites, several trips were
taken to a local kite shop, Powerline Sports in Seabrook, NH owned by Andrew Ghezzi,
where the proper use and control of large sport kites were demonstrated. These trips were
also used to find the kite line tension capabilities and the resulting coefficients of lift that
sport kites can generate. For smaller kites with less lifting capability a simple setup was
created:
Figure 14 - Beach Kite Testing
In this setup, the kite is attached to a digital tension meter that is then attached to
the ground. The readout of the digital tension meter allows us to find the line tension.
Since the four line kites have the ability to control angle of attack, several tension
measurements were taken for various angles of attack. A digital angle meter was used to
measure the angle θ in the diagram, which was used in turn to determine the kite’s
approximate angle of attack. However, one problem with our experimentation is that it is
nearly impossible to measure a kite’s angle of attack while it is in flight. To fix this
28
problem it was assumed that the angle of attack during the kites De-Powered Mode (0
line tension) is zero and that the angle of attack during Powered Mode (maximum Line
Tension) is 90 minus the angle theta that is measured at the ground.
Figure 15 - Angle of Attack Approximation
From the above diagram the angle of attack is approximated by:
(3)
This approximation is not always accurate because the parafoil could already be at
an angle when it is in the De-Powered mode. However, for our purposes the
approximation was accurate enough to determine a general form of several kites lift vs.
Angle of Attack ratios.
The one difficulty with this setup is that the digital tension meter was only
capable of measuring forces up to 60 pounds, when these kites were capable of lift well
over one hundred pounds. Therefore, for larger kites the digital tension meter was
replaced with a large industrial spring. By calibrating the spring to find its appropriate
spring constant, the kites lift was found by measuring the spring’s deflection.
29
3.2 System Design
A design process was conducted consisting of three steps to design the power
from kites demonstrator. First, we created a list of possible system designs. Next, each
design was evaluated until a final design mechanism was chosen based on several criteria.
Finally, the system was visualized using SolidWorks software.
3.2.1 Possible Mechanism Designs
To harness the available energy in an oscillating kite system, it is necessary to
design several mechanisms. The mechanisms that are required can be divided into three
separate categories; Angle of Attack Control, Energy Conversion, and Roll Control.
Though, the roll control mechanism was not completely designed in this project efforts
were made to address its potential design and function.
Angle of Attack Control
All of the potential energy conversion mechanisms require that the kite is
oscillating in a consistent up and down motion. To create this kite motion a mechanism is
required that will change the kites angle of attack. Modern kites used in the sport of kite
surfing have a standard setup that is used to control a kite’s angle of attack. Since these
modern angle of attack mechanisms are proven to operate successfully, it was decided
that it would be ideal to incorporate this existing technology into the attack change
mechanism.
30
The standard angle of attack control for a four-line sport kite is shown below:
Figure 16 - Modern Kite Control System in a De-Powered State
In the picture above, the inner line splits into two lines that are connected to the
leading edge of the kite while the outside two lines are connected to the kite’s trailing
edge. The inner two lines run through an opening in the control bar, allowing the bar to
slide along the two inner lines. In this type of control system the two inner lines provide
the major pull of the kite and the outer two lines connected to the control bar provide
angle of attack control. When the control bar is raised outward (Figure 16), the kite is in
the “De-powered” state, meaning that the Kite’s angle of attack is near zero and that the
kite is offering little or no pull. In this De-powered state the kite is effectively level and
all four lines connecting to the kite are at the same length. When the control bar is pulled
close to the rider as seen in Figure 17, the kite is in the “Powered” mode and the kite is at
full pulling force. In this instance, the trailing edge of the kite has been pulled downward
increasing the kite’s angle of attack. The two outer lines on the control bar can also be
‘trimmed’, in other words the line lengths can be shortened to a desired length by simply
31
tying knots in the cord that the lines attach to. The purpose of altering the length of the
lines is to optimize the kite’s L/D (lift/drag) and thus improve its overall performance.
When controlling the angle of attack of the kite, two significant conditions must
be considered. The upper and lower bounds of the control of the angle of attack are two
issues known as ‘undersheeting’ and ‘oversheeting’. Undersheeting is the condition at
which the kite is at an angle of attack providing an insufficient amount of lift. In this state
the drag is greater than the lift on the kite and therefore the kite is unable to lift and
produce significant power. Oversheeting is the condition in which the angle of attack has
increased beyond the angle of attack that coincides with the highest value of L/D. At this
point the kite begins to become less powered, and eventually depowered, as the angle of
attack increases beyond the maximum corresponding L/D.
Figure 17 - Control Mechanism in a "Powered" State
32
Energy conversion
The conversion of mechanical energy to electrical energy will require the use of a
mechanically driven electrical generator. However, there needs to be a mechanism that
can convert the oscillating motion of a kite into rotary motion that can power a generator.
Several possibilities for such a mechanism were considered.
The simplest conversion mechanism is a simple lever. A simple lever would be a
beam attached to a kite that would spin around a shaft powering an electrical generator.
Figure 18 - Simple Lever
The major advantage of this system is simplicity of construction. However, this
system was found to have several disadvantages. One major disadvantage is the system’s
inability to have the simple lever rotate in a complete circle. The shaft would require
enough momentum to keep revolving, while the kite would need to be put into “de-
33
powered” and “Powered” positions at key points in the revolution of the lever. This turns
out to be an extremely difficult task.
Another similar mechanism is a reversing simple lever. In this mechanism, the
kite would move the lever from side to side. The end of the lever attached to the
generator shaft would also be connected to a reversing clutch. This clutch would make
sure that the motion of the generator shaft only moves in one direction. An advantage of
this system is that the lever does not need to completely revolve around the entire
assembly. The angle of attack would change moving the lever from one side to the next.
However, a concern with this system is that it cannot sweep enough of an arc to have the
lever cause any significant torque upon a gear system.
Figure 19 - Simple Lever Reverse Clutch
Another possible mechanism consisted of a two-kite system. The two-kite system
involves two kites that work together to move the shaft of the generator. This can be
accomplished in several ways. Two kites could run the reversing simple lever mechanism
34
shown in Figure 19. Two kites could also be made to move in a circular motion that could
power a generator as in Figure 20. The difficulty of the two-kite system is the added
complexity. The kites could move around each other and twist together. It also becomes
more complex to control the motion of the kite. If the two kites aren’t configured
properly their motions could counterbalance each other.
Figure 20 - Possible Two-Kite System #1
Figure 21 - Possible Two kite System #2
35
There are also several classical mechanisms that have been used to convert
oscillating motion into rotary motion and vice versa. One of these mechanisms is the
pump jack. The pump jack is the mechanism used to pump oil from oil wells. An
example of a pump jack is shown in Figure 22.
Figure 22 - Oil Pump Jack
In the case of the pump jack, a motor creates a rotary motion the moves the left
side of the beam up and down, causing the right side of the upper beam to move up and
down in an oscillating motion. In the case of our project, the up and down motion of the
kite would move the right side of the pivoting beam up and down which would then drive
the beam that causes the rotary motion. This rotary motion would then drive the generator.
Other possible mechanisms involve the use of spring systems. This spring
systems would rely on the use of a sprag clutch. A sprag clutch is the mechanism that
36
allows a ratchet to apply a force in one rotational direction and not in the opposite
direction. An example of a possible system using a sprag clutch can be seen in Figure 23:
Figure 23 - Example Sprag Clutch
In a sprag clutch system, the kite would pull the flat rail gear that is positioned to
move up and down. When the kite is put into the power mode, the kite would lift the rail
upward, turning the large round gear and eventually spinning an electrical generator.
When the rail gear reaches a maximum height, the kite is triggered to change into “depower” mode. Once this happens the kite loses its upward pull and the spring pulls the
kite along with the rail gear back down into the starting position. As the gear rail moves
downward the gear moves back down but does not affect the gear shaft motion. When the
kite reaches the starting position, the kite is triggered into power mode and the cycle
starts all over again.
The last possibility considered was the combination between the sprag clutch and
the large rotating beam found on the pumpjack. The idea is that a kite would be attached
37
to an oscillating balanced beam. When the kite was in the down position it would
increase the kites angle of attack and the beam would be lifted upward by the kites
motion. When then beam reached the top of its arc, a sliding weight inside the rotating
beam would pull the angle of attack control strings downward, decreasing the kites angle
of attack. The weight within the beam would then pull the beam back down to starting
position. The Sprag clutch will come in to play when the beam is in its upward trajectory.
While the beam is moving upward it will be pulling a chain which is spinning a sprag
clutch that in turn leads to a gearbox and eventually a generator. A schematic of this is
shown here:
Figure 24 - Pumpjack/Sprag Clutch Combo
Roll Control
One possible problem with the kite motion is the kite’s ability to roll, or move
side to side. This problem must be dealt with in order to maintain the oscillating motion
of the kite and to prevent line tangling or even the total collapse of the kite in the air. In
38
order to control the roll of the kite, a simple mechanism can be integrated into the
construction of the overall system. An example of a simple mechanism that may control
the roll of the kite may be the addition of a spring system that will adjust for different
tensions in the lines of the kite. This type of system is shown here:
Figure 25 - Possible Roll Control Spring Mechanism
If the kite is farther to one side than the other the tension in the far side line will
be greater, and will in turn increase the spring force. A higher spring force will pull the
kite back to that side with a greater pull than the side with a lower force, ultimately
evening out the position of the kite in the air and preventing an excessive amount of roll.
As previously stated, the issue of roll control has been addressed but it has not been fully
developed nor has it been incorporated into the demonstrator design thus far.
3.2.2 Mechanisms Evaluation
After careful consideration of the possible design mechanisms, it was decided that
the two simple lever designs were the least likely to generate high amounts of power,
because they create low amounts of torque and rpm required to power a generator. These
two concepts were removed from the possibilities as well as any of the two-kite designs.
39
Two-kite designs are certainly feasible, but they are much too complex and costly for this
project.
With the elimination of a few of the possible design alternatives, the final choice
for the energy conversion mechanism system was limited to three different design
concepts. These three mechanisms were the Pumpjack, the Sprag Clutch and The
Pumpjack/Sprag Clutch Combo. To decide which mechanism provides the most benefit,
a comparable evaluation of all three mechanisms was conducted. The evaluation process
is described in the following section.
Figure 26 - Illustration of Pump Jack Design
40
Figure 27 - Illustration of Sprag Clutch Design
Figure 28 - Illustration of Sprag Clutch / Pump Jack Combo Design
Evaluation Scale
The scale in Table 2 was used to rate each system for a given criteria. The values
were assigned to each category with scores being awarded to each mechanism in
comparison to the other two mechanisms. In cases where the difference between two
mechanisms is indistinguishable, both mechanisms were awarded the same value.
41
Subjective
Good
Medium
Bad
Value
3
2
1
Description
Exhibits traits better than other mechanisms
Traits fall between two other systems
Exhibits traits worse than other two systems
Table 2 - Evaluation Scale
Evaluation Criteria:
1. Potential KW – The system with the potential to generate the most electricity. This
mechanism has the ability to harness the most mechanical power over a given time period.
The mechanisms ability to produce power at lower wind speeds was also considered.
2. Scalability – The potential for these systems to be scaled into larger systems. This
includes issues such as springs that lose their potential with size, and the functioning of
inner components of the system.
3. Practicality – The systems ability to be constructed and maintained. This includes the
availability and production methods included in the construction of the system.
4. Autonomy – this criterion assesses the systems ability to run without the aide of any
outside support. Criterion includes possible issues arising from in-climate weather and
varying wind speed.
5. Manufacturing – System ability to be constructed at low cost in a reasonable amount
of time. This includes complex parts that are already available or need to be specifically
machined.
6. Prototype cost – The predicted cost of each system based on the systems materials and
predicted times of construction.
7. Complexity – The systems complexity of design, including complexity of parts and
complexity of construction.
42
8. Variable Wind Speed – The mechanisms ability to operate over a large range of wind
speeds.
9. Demo Ease – ease of demonstration - A guess on which prototype would have the
highest chance of succeeding based on several evaluation categories already mentioned.
10. Going to Operation – The probability that this mechanism could be used in the rural
areas where farmers or local inhabitants could build this system for power generation.
This includes availability of construction materials and costs relative to other alternative
energy solutions.
11. Variability in Wind Direction – How well the mechanism will operate when the
wind direction changes from direction to another. This includes how the mechanism will
react when it is run with no supervision.
12. Stability Control – This assess the mechanism ability to recover after a large gust of
wind moves the kite in a direction that hinders the operation of the mechanism. Also
includes how easily a spring stabilizing system could be added to the mechanism.
Scoring Chart
Pump Jack
Sprag
Clutch/Tower
Pump Jack / Sprag
Clutch Combo
Potential KW
Potentially higher
output with desired
constant shaft
rotation Score: (3)
lower output due to
non-powered down
stroke, longer stroke
Score: (2)
Scalability
As size of beam is
increased potential
power is increased
Score: (3)
Relatively lower
output due to nonpowered down
stroke
Score: (1)
Too many variables
to consider in
scaling system
Score: (1)
Practicality
Issues with required
harmonic motion
Score: (1)
Similar to Pump
Jack, as size is
increased stroke
increases
Score: (2)
Complexity of AOA Removes technical
change mechanism issues from each
system
Score: (2)
43
Score: (3)
Autonomy
Manufacturing
After temp. stall,
unknown direction
or status of
operation of rotating
shaft
Score: (1)
Simple, but larger
components
Score: (2)
Prototype Cost
Net cost of
structure, e.g.
aluminum tubing
Score: (3)
Complexity
Simple design with
a complex motion
Score: (2)
Variable Wind
Speed
More variables that
have to change with
wind speed
(masses)
Score: (1)
Too many variables
to consider
Score: (1)
Demo Ease
Going to
Operation
Components of
system may be
readily available
Score: (2)
Variability in
Wind Direction
Different tether
angles may effect
the operation of the
AOA change
mechanism
Score: (2)
After temp. stall,
retraction
mechanism should
reset the kite cycle
Score: (2)
After temp. stall,
retraction
mechanism should
reset the kite cycle
Score: (3)
Simple and
relatively
inexpensive
components
Score: (3)
Relatively cheap
with surplus parts,
sprag clutch, control
bar
Score: (1)
Complex AOA
mechanism design,
simpler motion
Score: (1)
With spiral spring,
variable wind speed
is handled
Score: (3)
Simple components,
both large and small
Score: (1)
Aluminum tubing,
sprag clutch, gear,
chain
Score: ( 2)
Combo of the best
properties of each
mechanism
Score: (3)
Has spiral spring
yet the large arm
has masses to deal
with
Score: (2)
Complexity of AOA Complexity of
mech. may inhibit
tower and rotation
proper operation
eliminated, more
reliable operation
Score: (2)
Score: (3)
AOA mech. is a
Least complicated,
complex system
least # of issues,
making application could be used in
more complicated
simple applications
Score: (1)
Score: (3)
Kite tether comes
Different tether
out of tower
angles may effect
mechanism, keeping the operation of the
the movement in the AOA change
same direction no
mechanism
matter the angle of
Score: (2)
the line.
Score: (3)
44
Stability Control
TOTAL:
May be complex to
incorporate spring
system onto pump
jack arm
Score: (2)
23
Spring concept may
work best with
tower system
Score: (3)
23
May be complex to
incorporate spring
system onto pump
jack arm
Score: (2)
28
Table 3 - Evaluation Matrix
Qualitative Scale: Best mechanism (3), Second Best Mechanism (2), Third Mechanism (1)
Total Scores:
PumpJack: 23
Sprag Clutch: 23
Pumpjack/Sprag Clutch Combo: 28
Detailed Evaluation of Criteria
This section gives a detailed explanation on why values were assigned to a system
for each evaluation criterion.
Potential KW
The pumpjack scored the highest rating in this category because the pumpjack
provides energy to the generator in both the ascent and descent mode of the kite. The
other two mechanisms both received half of the pumpjack score because they only have
the potential to convert energy during the ascent stage, which is significantly less than the
pumpjack capabilities.
Scalability
The Pumpjack has the best scalability because it could potentially increase in size
as long as the kite providing the power was increased as well. However, both the Sprag
Pump and the combo are limited by the retraction mechanism that fails to operate
correctly when the system reaches a given size or dimension.
45
Practicality
The Pumpjack scored the lowest because it requires the most detail in design.
CAD simulations of this system show that it has problems running smoothly unless the
force of the kite is within a specific range. The Sprag clutch was second because of its
complex angle of attack mechanism. The Combo scored the highest because it removes
the technical issues from each of the other two mechanisms.
Autonomy
According to our modeling of the pumpjack system, it has the most difficulty in
running efficiently. The Pumpjack requires precise sizes and weights to effectively run
autonomously. This means that this system would require adjustments anytime the wind
shifted in speed.
Manufacturing
The Sprag clutch scored the best due to its ability to be manufactured in a
compact form with most of the component easily available. The Pumpjack was rated
second because of the larger components that would need to be machined in order to
function properly. The Combo scored last due to difficulties already mention in both the
Pumpjack and the Sprag Clutch.
Prototype Cost
The shear size of the Pumpjack makes it the most expensive unit. The pivoting
arm on the pumpjack system would need to be made of aluminum tubing to be
lightweight and withstand the various stresses applied during operation. The least
expensive system would be the sprag clutch. It requires the most complex building
46
procedure but it is the smallest system and requires the least amount of materials. The
combo scores in the middle because it comes both the inexpensive and expensive parts
Complexity
The Sprag Clutch is the most complex system because its angle of attack
mechanism has intricate mechanisms. The Pumpjack is the second most complicated
because the pumpjack needs to be fully adjustable to correctly operate in a full range of
wind speeds. The Combo scores the highest because it replaces the complex machinery
present in both the Pumpjack and the sprag clutch to easily operate in most conditions.
Variable Wind Speed
The sprag clutch rates the highest in this category because of its swiveling arm
capability. The swiveling arm allows the system to operate even when the wind is
changing direction. The Combo ranks second because the changes in wind direction may
affect the performance of the beam moving up and down. The Pumpjack scores last
because during simulations adjustments of wind direction on the run caused the pumpjack
to work inefficiently.
Demo Ease
The pumpjack is the hardest system for demonstration because of the number of
variables required for operation. The second hardest is the Sprag Clutch because of the
complex angle of attack mechanism. The simplest system is the Combo that is comprised
of the easiest parts of the two more complex mechanisms.
Going to Operation
The Combo is the least complex with the least number of issues. The Pumpjack
materials are the easiest to find, but it requires complex adjustability to become practical.
47
The sprag clutch is the least operational because it requires the more complex tower be
built, which is difficult for a rural area to produce.
Stability Control
The most stable would be the Sprag clutch because of its ability to point towards
the wind direction. The other two systems might not be affected in the same manner
because if the kite is pointing the wrong way the springs might not be able to properly
keep the kite stable.
Conclusion
Our conclusions show that the Sprag combo would be the most practical system
to build. Not enough decisive evidence was found to prove that the pumpjack system
would successfully operate at an effective capacity. The Sprag clutch would be the most
elegant design, but the Sprag combo would accomplish the same amount of energy
transfer with a simpler Angle of attack mechanism. In the next section we will discuss
detailed blueprints of our Sprag combo design.
3.2.3 System Design
Once the sprag combo design was chosen, a CAD schematic of the design was
created using SolidWorks software. The purpose of the SolidWorks schematic was not to
create a detailed blueprint of the system, but rather a framework for the system’s
construction. A screenshot of the SolidWorks design is shown in Figure 29.
48
Figure 29 - AutoCAD system Model
Figure 29 shows the initial design for the sprag combo system. The wooden
structure provides support to the rest of the system. The base structure was designed with
a wide base to prevent possible flipping from front to back during operation. A system
stress analysis provided in the results section shows that the structure can withstand the
potentially high stresses applied during operation.
The aluminum pivoting beam on the top of the structure is actually composed of
three separate beams. There is a large central beam with telescoping beams placed at each
end. An exploded view shows the beam in more detail.
49
Figure 30 - Telescoping Beam
The telescoping beams and the large central beam all have aligned holes drilled
through them. Bolts can be placed through the various holes to change the overall length
of the beam.
One essential portion of our design was the angle of attack mechanism. Our
original concept can be seen in Figure 31.
Figure 31 - Exploded View Angle of Attack Mechanism
50
The two lines extending from the top of the beam arm attach to the kite to connect
it to the arm and to allow for the angle of attack to change. The line on the right, as
shown in Figure 31, is a stationary line connecting to the ‘chicken loop’ of the kite
control bar. The line to the left is the line that will pull in and let out in order to change
the angle of attack of the kite. The left line is attached to the control bar, looped around a
smooth bolt, and finally connects to the weight inside the aluminum tube. When the beam
is lowered, the weight will slide to the position were the kite is changed to a high angle of
attack. This will cause the beam to rise upwards quickly. When the beam reaches the high
point in its swing, the weight slides and the kite is put into a low angle of attack. At this
point, the retraction springs overcome the force of the kite and pulls the beam back down.
When the beam is pulled back down, the kite is placed back into a high angle of attack
and the beam rises once again.
The most complicated portion of the structure is the gear system and retraction
springs. A close up of this portion of the system can be seen in Figure 32.
51
Figure 32 - Exploded View Gear System
As can be seen in this exploded view, the beam will pull on the chain that wraps
around the sprag clutch. The chain that wraps around the sprag clutch is then attached to
the chain channel. This channel in the pulley will hold the chain that is pulled in as the
retraction spring is recoiled. As the chain coils around the center of the pulley it layers on
top of itself which allows for a larger amount of chain to be pulled in with the increased
radius of the center of the coil. The channel is necessary because the beam pulls out a
chain length greater than the capacity of the largest commercial retractable springs
available. The gear ratio allows the torque and RPM of the final shaft of the generator to
be modified to provide the highest electrical output.
52
3.2.4 System Construction
The project group did the majority of the construction on the system. The
significant sections of the system are as follows: the beam to which the kite will be
attached; the gear system needed to turn the generator shaft; the angle of attack
mechanism; and the retraction mechanism that will be used to help pull the kite back
down to restart the cycle.
Main Wooden Structure
As stated in the previous section, the main structure of the system is a wooden
base with a form similar to that of a trebuchet. The base of the structure was constructed
with 4”x4” pressure treated lumber for a number of reasons. Lumber is very light in
comparison to metals such as steel and the pressure treatment will allow the structure to
withstand wet weather conditions. Also, the ease of construction played a major part in
the choice to use lumber as the system could be built anywhere with the proper tools and
no significant skilled labor, such as welding for steel, would be required.
The system was built with four legs that all stand at a sixty degree angle to the
ground for a wide base to maintain stability. Each of the tops of the legs was notched out
in order to connect to a horizontal piece of lumber intended to link the front and back
together as seen in Figure 33. At the bottom of the sides of the structure, braces were
installed in order to keep the legs separated at the desired sixty degree angle. The side
braces also have a second purpose as they will be the mounting points for much of the
retraction spring and gear systems as depicted in Figure 35.
53
Figure 33 - Top Side of Structure
Atop the structure are two lengths of lumber that span from side to side. The
pieces of lumber are once again 4”x4” stock with notches cut in the undersides at each
end in order to link these top pieces to the horizontal pieces spanning from the front legs
to the back legs. There are two purposes to these pieces of lumber. First of all they aid in
maintaining the stability of the structure when sideways shear forces are impinged on the
system. Secondly, these braces will provide a mounting point for the fulcrum point of the
beam arm as can been seen in Figure 33 and Figure 34.
Spanning from side to side on the faces of the front and back legs are 2 more sets
of braces, as seen in Figure 35. The first pair of braces is attached to the faces of the legs
at the bottom of the structure just underneath the braces that span the sides. The second
set of front braces are attached approximately three-fourths of the way up the faces of the
legs. These two pairs of braces are intended to maintain the rigidity of the structure as
well as maintain the desired span between the sides of the structure.
54
Figure 34 - Structure Braces and Beam Fulcrum Point
Figure 35 - Frontal View of Structure
Turnbuckles
When the main wooden structure was first constructed instability in the structure
was observed when a force was applied at the top of the structure from side to side. In
order to stiffen the structure a system of turnbuckles were used to apply tension on each
side of the structure in an “X” pattern. The turnbuckle assemblies consist of several parts
and are as follows:
55
•
•
•
•
1/8” steel wire
Steel turnbuckles – open hook / eyelet
Steel wire loops and clamps
3/8” eyelet screws
The eyelet screws were screwed into the faces of the front and back legs just
above the bottom braces and below the top braces, as seen in Figure 36, as the main
mounting points for the turnbuckle assemblies. The open ends of the turnbuckles were
hooked onto the eyelet screws. Steel wire loops were hooked onto the closed end of each
of the turnbuckles and the steel wire was threaded from one turnbuckle to the opposing
turnbuckle to form the desired “X” pattern. Once the steel wire was threaded the clamps
were attached to the wire to secure the wire in place and the turnbuckles were tightened
to apply the desired tension on the structure. With the addition of the turnbuckles, the
structure became a completely rigid structure when considering an applied side – to –
side force.
Figure 36 - Turnbuckle Assembly
56
Balanced Arm Assembly
In the final design of the kite power demonstrator the kite will attach to the top of
the beam at the top of the structure. In order to maximize the number of different testing
setups, the balanced beam was designed with two main features. First, the beam is
constructed of two different sizes of aluminum square tubing in order to allow for
telescoping of the smaller tube inside the large tube. The beam consists of two square
aluminum tubes with outer diameter of 3.5” at a length of 5’ each and a third length with
an inner diameter of 3.615” and a length of 8’. Second, a series of holes was drilled in the
sides, top, and bottom of the 3 lengths of aluminum tubing in order to secure the lengths
of aluminum at any desired length for testing purposes. These length changes will allow
for changes in weight and beam length from the front of the structure to the back.
Figure 37 - Balanced Arm Assembly
57
Fulcrum Point Assembly
The fulcrum point is the attachment point at the top of the wooden structure for
the aluminum beam. In order to enhance the ability of the beam to change in length, as
discussed above, the beam is set inside another square tube with holes drilled into it in the
same fashion as the beam. Therefore, the beam can now attach at the fulcrum point with
bolts in different holes. The steel square tube has an outer diameter of 5” and in inner
diameter of 4.5”. There is a significant amount of variance between the outer diameter of
the large aluminum tube (4”) and the inner diameter of the steel tube in order to have the
ability to thread bolts down through the bottom of the steel tube.
The bolts that were threaded through the bottom of the steel tube are attached to
two pillow block bearings with an inner bearing diameter of 1”. On top of the structure
there are an additional two pillow blocks mounted to the top wooden braces spanning the
structure. These two additional pillow blocks make up the second half of the rotating
fulcrum point with the addition of a 1” diameter steel rod. The steel rod is threaded
through the four pillow blocks to form the axle for the fulcrum point. Figure 38 shows the
fully assembled fulcrum point.
Figure 38 - Beam fulcrum point assembly
58
Angle of Attack Change Mechanism Assembly
One of the most essential features of the kite power demonstrator is the assembly
of the angle of attack change mechanism. When determining which system design would
work best for this project one key consideration was the ability to incorporate a
mechanism to change the angle of attack of the kite. In the final design of the system the
angle of attack change mechanism was incorporated into the balanced beam arm
described above.
To limit the complexity of the system it was essential to design a very novel AOA
change mechanism. The final design incorporates the inner square tube at the front of the
demonstrator and a simple weight system. As shown in Figure 39, there is a series of
holes drilled through the sides of the small aluminum tube. These drilled holes are
intended to hold two bolts, which are currently inserted into the holes at the very far right
of the arm, with the purpose of stopping the sliding weight of the mechanism. A range of
holes was drilled in order to apply another level of testing in the system design. The
current design enables the bolts that stop the weight to be either ahead of or behind the
kite attachment point.
The attachment point for the kite is the eyelet bolt located on the arm in the left
side of Figure 39. This bolt will be utilized by attaching a carabineer to the closed loop of
the bolt and then attaching the ‘chicken loop’ of the kite control bar to the carabineer.
This setup will attach the kite to the beam arm in order to lift it during motion. The bolt,
only hex nut visible, to the left of the eyelet bolt is the point around which the line from
the weight to the control bar will pass as it is pulled through a hole in the top of the beam
59
arm. Rather than using a pulley deep inside the beam arm it was believed to be a more
novel design to simply use a bolt with a smooth surface as the turning point for the line.
Figure 39 - Angle of Attack Change Mechanism Assembly
Axles / Gears Assembly
The portion of the system intended to convert the linear motion into a rotary
motion consists of a series of axles and gears as shown in Figure 32. The gear assembly
consists of two steel axles of 1” diameter, two 9.2” sprockets, one 2.3” sprocket, and four
pillow block bearings. At the front of the system the main axle is mounted with two
pillow blocks attached to the wooden side brace that extends out from the front of the
structure. On this axle, the two large sprockets are mounted. As viewed in Figure 40, the
left large sprocket is intended to be used with the retraction mechanism and the right
sprocket is part of a ‘gearbox’ system. The axle mounted to the back of the front structure
legs is the second half of the ‘gearbox’ system. This axle is also mounted on two pillow
blocks and holds the small sprocket. In order to turn the small sprocket a chain will be
looped around both the large and small sprockets. The ratio in sizes from the large
60
sprocket to the small sprocket will allow for a greater number of rotations of the small
sprocket, increasing the rpm value and thus the potential power output.
The third axle shown in Figure 40, resting on the brace in front of the mounted
sprockets, is intended for use with the retraction spring system. This axle would also be
mounted with two pillow blocks, but this mechanism has not been completely designed
and will require future work.
Figure 40 - Axles and Gears Assembly
3.3 System Simulations
The ultimate goal of mechanism simulation is to predict the kite motion and the
power generated by the mechanism. The two major figures of interest are the RPM of the
shaft and the Torque it is experiencing. Once initial estimates of these quantities are
found, it then will then be possible to manipulate the simulations to find the most
efficient design of the system. From these most efficient designs we can also predict the
eventual power producing capabilities of our mechanism. Simulations for the system
61
were based on both steady state simulations and dynamic model involving differential
equations.
3.3.1 Steady State Simulations
The steady state kite modeling we conducted is based upon previous work. Some
important variables are defined in Figure 41:
Figure 41 - Kite Model, From Goela (1983)
Goela’s steady state analysis is coupled to a model of the oscillating arm in our
demonstrator. In addition, a new aerodynamic model is incorporated to better model the
kite behavior. The derivation of the entire model begins with the motion of the kite.
Variables used in the steady state kite calculations are:
62
The goal of the steady state kite calculations is to find the kite tether tension
created by the kite. From Goela (1986), it was shown that the normalized kite velocity,
normalized kite weight, and the lift over drag ratios are required to develop an equation
for the kite tether tension.
The first step towards the normalized kite velocity is determining the velocity of
the kite given by:
(4)
Where the Velocity of the Kite from arm motion is calculated using:
(5)
Two of these terms come from the design of the Mechanism:
With the total kite velocity we can then find a value for the kite Normalized Velocity:
(6)
For the lift over drag coefficients, equations modeling the aerodynamic
performance of airfoils were added to Goela’s (1983) steady state analysis. This
aerodynamic model replaced the L/D values used in Goela (1983). A range of angles of
attack were simulated through the use of Lift over Drag ratios that simulate the kite at
various angles of attack. This range of L/D values can be determined from basic airfoil
theory and finite wing theory. Using this theory we can approximate values of Cl and Cd
that define L/D. The Coefficient of Lift (Cl) is found using basic airfoil theory:
63
(7)
Using finite wing theory the derivative of the coefficient of lift is assumed to equal:
(8)
Substituting equation (8) into Equation (7) gives the coefficient of lift:
(9)
The coefficient of drag (Cd) is found by:
(10)
Using the calculated values of Cl and Cd we can find a range of Lift over Drag ratios
with:
(11)
The inverse of Lift over Drag is simply:
(12)
Equations (11) and (12) represent Lift over drag and drag over lift coefficients
that are functions of the kite’s angle of attack (α). The last step is to find values for
normalized kite weight:
64
(13)
Finally, from Goela (1983) the calculation of a line Tension Normalized
coefficient is found:
(14)
From this normalized equation we can find numerical tension values:
(15)
Using the line tension a formula can be found to determine the power harnessed
by the oscillating beam in the form:
(16)
Using the Line tension from the steady state kite model, we can estimate the
forces acting upon our mechanism design. Additional parameters used to model the
oscillating beam design were included. Steady state equations for the mechanism were
based on the model seen in the following figure:
65
Figure 42 - Pivoting Arm Diagram
Variables for the pivoting arm simulation:
With the line tension calculated from steady state theory, it is possible to sum
moment’s about the rotating arm’s pivot point to calculate the tension that would be
created on the chain going into the gear system, as follows:
(17)
Solving for the Chain tension yields:
66
(18)
Using calculated values for the kite line tension, the chain tension can now be
determined. The next step is then to use this chain tension to estimate Torque and
rotational speed of a generator shaft. A simple gear system shown in the Figure 43:
Figure 43 - Two Gear-System Schematic
In the gear system shown above, the chain attached to the pivoting beam is
spinning the shaft on gear #1, and is giving Gear #1 a given torque and RPM value. The
following values are defined for the gear system:
Torque on a given gear is usually approximated by multiplying the chain tension
and the radius of the gear or the gear ratio of the gearbox in use. From this approximation
we can find the torque on the generator shaft using the following equation:
(19)
67
The predicted RPM of the generator shaft depends Stroke time of the rotating arm,
which is approximated by finding the time required for a kite to travel through its power
phase. The RPM of Gear #1 and Gear #2 are found using:
(20)
Using the value of this rpm we can estimate the RPM this would create on the
connected Gear #2:
(21)
This is the final step in our steady state theory. Results from the steady state
theory model will be given in a later section.
3.3.2 Dynamic Modeling Simulation
Dynamic simulations of the system provide a much deeper look at the possible
power production and overall operation of the entire mechanism. This simulation allows
the operational characteristics of the system to be viewed in response to time. It also
allows for average values of system production that can be used to determine valuable
information. The major advantage of this system over the steady state theory is that it
allows predictions on the functionality of the system. The system will have the ability to
predict whether a given mechanism setup will be able to run effectively or autonomously.
For the dynamic modeling, the simulation was created to model a slightly
modified version of the mechanism. This slightly modified version of the system
included a load on one end of the bar to replace the downward pull of the retraction
68
spring. This was created due to worries that a retractable spring would not be found in
time for the systems construction. Therefore the system was simulated as if our project
team needed an extra load to prove the concept that the pivoting beam could continually
rock in an up and down movement. A diagram of the modified mechanism is shown here:
Figure 44 - Modified System Simulation
One difference with this simulation is that the overall power of the system will be
based off of the mechanical power generated by the rotating arm. This power is
approximated using:
(22)
Where:
The mathematical equations used for this dynamic simulation begin with kite
dynamics found in Dr. Goela’s equations. Using Figure 45, two coordinates are defined:
69
Figure 45 - Simulation Parameters
In this figure V1a represents the velocity of the kite in the tether line direction
while V2a represents the kite velocity perpendicular to the kite tether direction. From
Goela (1986), we find an equation for the change in velocity in the V2a direction:
(23)
Where the end term is the coriollis force based on a variable tether line length
used in the Goela (1986) report:
(24)
70
The sprag pump has no variation in the tether line length. Since the kite is
attached to the beam it is obvious that there is no velocity or movement towards the V1a
direction. Therefore the coriollis force can be ignored because the velocity in the V1a
direction is:
(25)
Therefore the change in velocity of the kite in the V2a direction is:
(26)
Looking in the V1a direction it is found from Goela (1986) that the change in the
kite’s velocity is assumed to be:
(27)
Since the V1a derivative is known to have a value equal to zero, the equation can
be rearranged to solve for the kite tension. The kite tension is now found as:
(28)
Using the equations already mentioned the dynamics of the kite can be simulated
based on a given set of initial conditions for the kite. The next step is to couple the kite
dynamics into the oscillating motion of the arm. The change in the angle theta is
calculated based upon the movement of the kite’s velocity. The velocity in the V2a
direction shows:
(29)
This can be rearranged to express:
71
(30)
Further calculations coupling the kite motion and the oscillating arm can are
found using another figure:
Figure 46 - Simulation Parameters Part #2
In this figure, Va represents the velocity of the beam perpendicular to the beam
while Vat represents the velocity of the beam parallel to the beams orientation. In this
case the sum of the moments about the rotating beam culminates in the following
equation:
(31)
In this case it is very obvious that the beam does not have any motion in the Vat
direction because of the beam arm is fixed in a rotating form. Therefore it can be
assumed that:
72
The velocity in the Va direction can be calculated from:
(32)
Where the angular velocity of the beam is:
(33)
Hence, the following equation if found:
(34)
The power of the mechanical system is calculated based on the power of the
rotating beam. This is calculated by:
(35)
Substituting for the angular velocity we find that the power is:
(36)
The main key to the simulations lie in the four differential equations (23), (30),
(31), and (34). Using these four equations and setting some initial quantities allows for
the determination of the kite and oscillating beam dynamics. With Matlab, a given time
step is used to determine the number of iterations involved in the differential equations.
Once initial quantities are set, the Matlab program runs the coupled equations over a set
time limit to simulate the system dynamics.
One problem with these equations is that they do not involve the use of the angle
of attack mechanism. They only describe the beam and kite motion as it moves in one
direction, either up or down. Therefore, Matlab is set to a point that when the oscillating
73
beam reaches a given height, the kite’s angle of attack is altered. When the beam reached
a given angle with the horizontal field, the code changes the kite’s angle of attack to
respond to the changing beam angle with the horizontal. For example, when the rotating
mathematically reaches a given downward angle with the horizontal, the kite’s angle of
attack is changed to a higher set value that triggers the kite to provide a force. An
example is shown here:
Figure 47 - Rotating Beam reaches a set Angle in the program
Results from this dynamic model simulation are given in a later section.
3.4 System Stress Analysis
Stress calculations for the Sprag Combo demonstrator begin with the analysis of
the long pivoting beam. The concern is whether the beam can withstand the full force of
the kite pulling on one end of the beam. The stress calculations for the beam are based on
the equation of a beam fixed at one end with a load placed at the free end.
74
Figure 48 - Beam Stress
The highest stress that the beam will be placed under is at the fixed end support of
the beam. The stress at this point can be calculated from the equation:
(37)
Where the terms in this equation are identified as:
The beam section modulus is found by analyzing the cross section of the beam:
The other concern with the beam is its deflection. The largest deflection that the
beam will undergo is at the free end where the kite force is located. This maximum
deflection is found using the equation:
75
(40)
The new term introduced in this equation is the Elasticity Module (E) that is
determined by the beam material. The next analyzed component of the system is the
center beam on which the pivot rests (Figure 49).
Figure 49 - Top Support Beam
The beam can be shown as a beam supported at two ends with a downward force
in the middle. A body force diagram shows:
Figure 50 - Beam Stress Diagram
76
From this type of load on the beam the stress at the center of the constant cross
section is determined from the equation:
(41)
The maximum deflection of the center beam is at the center of the beam at the
pivots location and is calculated using:
(42)
These components represent the major concern for the stress analysis of our structure.
77
4. Results
Results come directly from the equations and details explained in our
methodology section. They are divided into three separate sections. The first of these
sections is the data collected from physical testing of the kite systems. The second and
third sections are the numerical predictions of the model simulations and the possible
stresses that the model simulation predictions will create on our mechanical system.
4.1 Kite Testing
Several trips were made to the local sport kite shop, PowerLine Sports in
Seabrook, NH, where three kites were tested. Two of these kites were demonstration kites
that were tested using our digital tension meter setup. The first of these kites was the
Ozone Axis 4 m^2. The kite and its characteristics are listed below:
Figure 51 - Ozone Access 4m
Kite Area (m^2)
Kite Weight (Kg)
Kite Length
2 m^2
1 lb. 5 oz.
135 (in.)
78
Middle Width
Tip Width
# of Control Lines
Minimum Required Wind Speed (m/s)
46 (in.)
30 (in.)
Four
3 m/s
Table 4 - Ozone Axis Kit Characteristics
Using this kite the following data was collected:
Average Wind Speed
(mph)
Angle of Attack
(deg.)
Tension Reading
(lb)
Coefficients of
Lift
De-Powered
avg. 10.5 mph
avg. 70-74
4
7
8
11
13
Avg 0-11
0.33
0.58
0.66
0.90
1.07
avg. 10.5 mph
80
80
85
70
60
avg. 70-80
7
4
11
20
20
avg. 12-20
0.575743118
0.328996067
0.904739185
1.644980336
1.644980336
Powered
Table 5 - Ozone Axis Tension Data
From our data collection, an assessment for the kite’s coefficient of lift data was created.
From the data, at assumed zero angle attack the average Cl was:
At α = 18 degrees the average lift was 16 lb. resulting in a Cl of:
79
Since the coefficient of lift vs. Angle of attack is linear we can approximate the
slope of the line between α = 0 and α = 18. This slope was found to be:
Hence an approximate equation for the kite’s coefficient of lift as a function of
angle of attack:
(43)
The second kite tested was the 7 m^2 Cabrinha Crossbow. For this kite, the kite
line tension was measured along with the force required to move the Power/De-power
bar. The force on the Power/De-Power bar is important because it gives an idea for the
size or size of a counterweight or spring mechanism. The Cabrinha kite and its
characteristics are detailed below:
Figure 52 - Cabrinha Crossbow 7m
Kite Area (m^2)
Kite Weight (Kg)
Kite Length
Middle Width
7 m^2
5 lb. 4 oz.
215 (in.)
55 (in.)
80
Last strut Width
# of Control Lines
Minimum Required Wind Speed (m/s)
38 (in.)
Four
3 m/s
Table 6 - Cabrihna Cross Kite Characteristics
The Cabrihna Crossbow is an LEI kite. The inflatable edge and inflatable struts
are filled with air that is pumped into the kite through a hand air pump. For this kite the
digital tension meter setup was also used and the following data was collected:
Average Wind Speed
(mph)
Angle of Attack
(deg.)
Tension Reading
(lb)
Coefficients of Lift
Depowered
avg. 13.5 mph
avg. 79-85
Avg. 0-4
0.05686621
avg. 13.5 mph
avg. 79-85
42
28
22
24
23
22
16
10
12
38
23
16
18
21
24
22
18
35
20
mostly stayed at
23
1.194190408
0.796126939
0.625528309
0.682394519
0.653961414
0.625528309
0.454929679
0.28433105
0.341197259
1.080457988
0.653961414
0.454929679
0.511795889
0.597095204
0.682394519
0.625528309
0.511795889
0.995158674
0.568662099
Powered
0.653961414
Table 7 - Collected Data for Cabrihna Crossbow 7 m^2
The angle of attack is approximated in these calculations in the same way that it
was approximated in the calculations for the Ozone Axis. At assumed zero angle of
attack the average Cl was:
81
At α=8 deg. the average lift was 23 lb. resulting in a Cl of:
The slope of the line between α=0 and α=8 is found to be:
An equation can be found for the coefficient of lift as a function of Angle of Attack:
(44)
Additionally, the force required on the Power/de-power bar was measured to vary
from 0-10 lb., with the force becoming greater as the bar was moved into the powered
position.
The third and final kite tested was the Peter Lynn Guerilla that was purchased as
part of the project. The kite was larger then the other two previously tested kites, and it
was calculated that the kite would pull with a force greater then the capability of the
tension meter. Therefore, we used a spring to determine the spring tension. The spring
used was a large industrial spring found in Washburn Laboratories.
The first step in using the spring was finding its calibration curve so that we could
determine the length that the string stretches for a given force. In our laboratory, the
spring’s extension was measured for a variety of different weights. The data taken is seen
here:
82
Weight (N)
145.7375
179.66875
282.01875
348.76875
460.01875
571.26875
Spring Extension (in.)
0.625
1.0625
2.375
3.25
4.875
7
A graph of the data should represent a linear line to represent a spring deflection
vs. a load. Plotting the data shows:
Spring Deflection (in.) vs. Weight (N)
8
Deflection (in.)
7
6
5
4
3
2
1
0
0
100
200
300
400
500
600
Weight (N)
Figure 53 - Spring Deflection vs. Weight (N)
Using this chart the deflection of the spring was measured with the movement of
the kite. This spring method testing our kite was attempted once but due to bad weather
conditions at the time of testing, unreliable results were obtained.
83
4.2 Determination of Simulation Values
From the data acquired from our beach testing, and the known values concerning
our kite, it is possible to make predictions based on the steady state theory, pivoting arm,
and RPM/Torque based on the simulation models.
4.2.1 Steady Sate Simulation Results
Beginning with the steady state kite theory, the known and assumed values are
determined. The following values are known:
From direct observations it has been determined that the approximate value can be
assumed for the kite motion:
The rotational speed of the oscillating arm is approximated based upon a
reasonable estimation:
Using the known and unknown values, the first calculation is the velocity of the
kite due to the arm motion. From equation (5), this value is calculated:
Then using equation (4), the velocity of the kite is:
84
The normalized velocity of the kite is:
With this in mind, the coefficients of lift and the corresponding normal weight of
the kite values are made to be functions of a range of angles of attack. Using these values
the tether line tension is calculated as a function of angle of attack:
Figure 54 - Kite Line Tension vs. Lift over Drag
From this graph we can estimate that our system will have a line tension force
varying anywhere from 75 to 400 Newton’s. Now that we have discovered our line
tension we can use these values to find the possible chain tension using our pivoting arm
model. For the pivoting arm model we used the following values:
85
Using equation (16) we find the calculated power of the oscillating beam as a
function of Angle of Attack:
Figure 55 - Pivoting Beam Power vs. Angle of Attack
Using equation (18) we find the tension on the chain to vary as a function of
Angle of Attack:
86
Figure 56 - Chain Tension vs. Lift over Drag
From this graph we can see that estimates show the chain tension to vary from 150 to a possible 200 N of force. The negative values represent that the weight of the
pivoting beam arm is still greater than the force of the kite. It can be seen that we can’t
expect the beam to move until the kite reaches a certain degree for its angle of attack.
Finally, this predicted chain force can then be applied to our gear ratio modeling
to determine the possible end torque value for a gear system. For this portion we assume
gear that will be used on our gear system. Therefore, our known values are:
87
From these values we can determine the final torque on the generator shaft as:
Figure 57 - Torque vs. Lift over Drag
Similar to the Chain tension, the beginning torque is negative when the beam the
kite is not moving the beam upward. When the beam finally moves upward we see that
the torque on the generator approaches 6 N-m. Using the given values, the possible final
RPM on a generator shaft can also be estimated as:
88
4.2.2 Dynamic Simulation Results
In this section we present results from the dynamic simulations of the kite and
rotating arm system. The equations of motion solved in the simulation were summarized
in Section 3.3.2.
For the mechanism with the counterweight system, it is ideal to determine what
counterweight would produce the highest average coefficient of power for a given wind
speed. The Dynamic simulation in the Matlab program can solve for coefficients of
power in terms of the counterweights for varying wind speeds. The results for different
counterweights and wind speeds runs are presented in Figure 58:
CpAvg. vs. WCTR
0.12
0.1
3 m/s
CpAvg.
0.08
5 m/s
7m/s
9 m/s
11 m/s
13 m/s
15 m/s
0.06
0.04
0.02
0
0
500
1000
1500
2000
WCTR (N)
Figure 58 - Average Coefficient of Power vs. Counterweight (N)
89
From Figure 59, it is apparent that an optimal counterweight exists for each wind
velocity. In Figure 59, these optimal WCTR values are presented:
WCTR vs. Wind Speed (Optimal Average Power)
1800
3
2
y = 0.8854x - 13.103x + 117.86x - 222.08
1600
1400
WCTR (N)
1200
1000
Series1
Poly. (Series1)
800
600
400
200
0
0
5
10
15
20
Wind Speed (m/s)
Figure 59 - Counterweight of highest Power Production
Figure 59 could be used during testing of the demonstrator. An operator could
reference this chart to adjust the systems counterweight to provide the highest power
production for a given wind speed. A curve fit to Figure 59 yields:
(45)
This power coefficient, defined in equation (), is a measure of the fraction of available
power available in the wind that is extracted by the mechanism.
(46)
In Figure 60, the power coefficient is plotted for the optimal condition.
90
CpAvg (Ideal) vs. Wind Speed (m/s)
0.12
y = -2E-05x 4 + 0.0007x 3 - 0.0109x 2 + 0.0675x - 0.0499
CpAvg (Ideal)
0.1
0.08
Series1
0.06
Poly. (Series1)
0.04
0.02
0
0
2
4
6
8
10
12
14
16
Wind Speed (m/s)
Figure 60 - Highest Coefficients of Power for Wind Speed
From Figure 60, it can be seen that the maximum coefficients of power are
reached during wind speeds between five to seven meters per second. This is ideal for the
purposes of the project because these wind speeds are the average wind speeds that our
kite system will be experience in an area such as Worcester. A curve fit to Figure 60
yields:
(47)
From the average coefficients of power, we can find the optimal power
production that the system will generate for a given wind speed. A graph of power
produced versus wind speed is given in Figure 61:
91
Power (KW) vs. Wind Speed (m/s)
2.5
y = -0.0013x 3 + 0.0449x 2 - 0.2666x + 0.4688
Power (KW)
2
1.5
Series1
Poly. (Series1)
1
0.5
0
0
2
4
6
8
10
12
14
16
Wind Speed (m/s)
Figure 61 - Optimal Power Production for given wind speed
A curve fit to Figure 61 yields:
(48)
The results of this data show that the kite system will have the ability to produce
power approaching the production of similar sized turbine systems. The power
Production for our kite system at an average wind speed of 8 m/s is around 0.5 KW. A
similar sized turbine at this wind speed and power production would have a rotor
diameter of 4.652 meters. For New England areas, such as Worcester, the average wind
speed at a high altitude would generate around half a kilowatt.
The average yearly power production for each wind speed is presented in Figure
62. This term is very common in the wind industry and is often used as benchmark of
comparison for electrical generating devices.
92
Power (KW*h/year) vs. Average Wind Speed (m/s)
25000
y = -11.324x 3 + 393.55x 2 - 2335x + 4106.6
Power (KW*h/year)
20000
15000
Series1
Poly. (Series1)
10000
5000
0
0
2
4
6
8
10
12
14
16
Wind Speed (m/s)
Figure 62 - Yearly power production vs. average wind speed
A curve fit to Figure 62 yields:
(49)
A further advantage of the dynamic system simulation is the ability to predict the
kite’s motion. Using the Matlab software we can simulate the path of the kite as it moves
through the sky as shown in Figure 63. The wind velocity was assumed to be at 5 m/s,
with a counterweight of 135 Newton’s set to provide the highest possible average
coefficient of power.
93
Figure 63 - Kite Motion
The starting point of the kite is at the origin (0,0). Figure () shows that over time,
the kite oscillates in a repeating motion. Region A represents when the beam is lowered
and the kite’s angle of attack has been changed. In this region the growing force of the
kite is decelerating the beam. Region B represents the area where the kite and beam are
rising upwards. This is the power stroke phase of the system. Region C represents the depowering of the kite combined with the oscillating beam falling downward.
In Figure 64 the oscillating motion of the beam is presented. The angle of the
beam over time is shown as:
94
Angle Gamma (deg.) vs. Time(s)
60
40
Gamma (deg.
20
0
0
5
10
15
20
25
30
35
-20
-40
-60
-80
Time (s)
Figure 64 - Simulated Angle gamma (deg.) vs. Time (s)
The oscillation of the beam appears to oscillate between:
When θ = 0, the oscillating beam is the horizontal. At θ = 40 is the beam is 40
degrees above the horizontal plane and at θ = -60 the beam is 60 degrees below the
horizontal plane. The oscillation in the oscillating beam corresponds to the change in the
kite’s angle of attack. When the Gamma angle is at -60 degrees the kite’s angle of attack
reaches a maximum, where it then pulls the beam upwards increasing the beam’s Gamma
angle. As the Gamma angle increases, the kite’s angle of attack decreases slowly once
theta is greater than the horizontal. Eventually at 40 degrees the Kites angle of attack has
reached zero and when of the beam pulls it back down where the process repeats.
95
We can also determine the varying line tension caused by the kite’s motion:
Line Tension (N) vs. Time (s)
1600
1400
Line Tension (N)
1200
1000
800
Series1
600
400
200
0
-200
0
5
10
15
20
25
30
35
Time (s)
Figure 65 - Simulated Line Tension (N) vs. Time (s)
In Figure 65 it is interesting to note the periodic large amplitude peaks. These
high peaks occur when the kite suddenly changes its angle of attack as the rotary beam
moves downward in the descent phase. The large amplitude line tension serves to
decelerate the rotary beam as it descends. These large line tensions could fracture the kite
tethers if the line tension exceeds the tether tensile strength. We estimate the tensile
strength of each of the kite line at 2648 Newton’s. We estimate the majority of the line
tension to reside on two kite lines, yielding a margin of safety around 3.5. It is also
important to realize that the dynamic simulation alters the Angle of Attack between the
ascent and descent stages over a 0.25 second time interval. If the time is larger in the
actual operation of the demonstrator, these peak line tensions may be reduced.
In Figure 66, the Lift over Drag predictions from the added aerodynamic analysis
are presented:
96
LD vs. Time (s)
8
6
4
LD
2
0
0
5
10
15
20
25
30
35
-2
-4
-6
Time (s)
Figure 66 - Simulated Lift over Drag vs. Time (s)
The downward large amplitude peaks in the lift over drag ratio are a result of the
high peaks in the line tension. The maximum L/D values are around 6 when the kite and
beam are moving during the ascent phase. The minimum L/D varies between one and
four during the descent stages of the kite’s motion. These L/D values reasonably
approximate the L/D values set by Goela (1986).
In Figure 67, the values of the coefficients of lift versus the coefficients of drag
are presented:
97
Cl vs. Cd
2.5
2
1.5
Cl
1
0.5
0
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
-0.5
-1
Cd
Figure 67 - Simulated Cl vs. Cd
4.3 System Stress Analysis
Stress calculations for the Sprag Combo begin with the analysis of the long
pivoting beam. The concern is whether the beam can withstand the full force of the kite
pulling on one end of the beam. The stress calculations for the beam are based on the
equations found in Section 3.4. For the aluminum beam being used in the system the
cross section lengths of the beam and the calculated moments of inertia and section
modulus are:
For the stress calculations a beam length of approximately 7 feet was assumed,
while the kite force was approximated from our calculations in the system simulation
section. Using this variable kite tension from steady state analysis, the shear stress on the
beam as it moves upwards can be determined. The final calculations show:
98
Figure 68 - Stress vs. AOA
The calculations show that the beam stress will vary between from 1-8(Ksi),
which is well below the yield strength of aluminum (32 Ksi). The other concern with the
beam is its deflection. The beam material is Aluminum 6063, with an Elasticity module
of:
Using equation (40) and kite forces from steady state, Figure 69 shows the Beam
Deflection vs. Angle of Attack.
99
Figure 69 - Deflection vs. AOA
Calculations for the beam deflection show that the beam will deflect from 0-3
inches over the course of the kite’s ascent.
The next questionable portion of the system is the center beam on which the pivot
rests. The beam in this case is a 4x4 cross section of wood. Wood is considered to have
an Elasticity module of approximately 10MPa. The Moment of Inertia and section
modulus for the beam is calculated to be:
The load placed at the pivot point is effectively equal to the moment created on
the support beam of created by the conjunction of the oscillating arm and the kite force.
Using the moments created by the oscillating beam, the stress on the center beam is found
to vary from 1-4ksi. Well within the yield strength of wood.
100
Figure 70 - Support Beam vs. Kite Angle of Attack
The maximum deflection of the support beam is at the center of the beam near the
pivot of the oscillating beam. Using equation (42), the deflection is:
Figure 71 - Beam Deflection vs. Angle of Attack
From Figure 71, the beam deflection on the central support beam is minimal.
101
5. Conclusions
To begin the process of designing the kite power demonstrator different types of
kites were researched until a final decision could be made as to the ideal kite for a
demonstrator. The chosen kite is any one of a series of production kite boarding kites
known as twinskin kites manufactured by Peter Lynn. The main benefit of these kites was
the auto-zenith feature that enabled the kite to remain stable in the air with minimal user
input.
Two different testing setups were completed in the overall kite testing portion of
this project. Preliminary testing of kite boarding kites was done with several different
kites with the goal of measuring potential line tensions, though none of these kites were a
twinskin model. After the purchase of the Peter Lynn Guerilla twinskin kite-boarding kite,
testing proceeded to determine potential line tensions to determine kite lifts and other key
kite dynamic characteristics for simulation purposes.
The second portion of the project was to design and build the kite power
demonstrator to be used in concert with the Peter Lynn Guerilla. Several different
mechanism concepts were considered until an evaluation matrix was used to determine
the best option for the demonstrator. After the decision for the basic design was made,
design modifications and optimizations were made in order to simplify the system until
the design iterations led to the final design presented in this project report. Construction
commenced with the completion of the system design. The current stage of construction
is discussed in section 3.2.4. At this stage, the construction to be completed pertains to
the retraction spring which will need some future work before completion.
102
Tying together the two main aspects of the project, the kite and the mechanism,
are the various simulations completed. Simulations began with simple steady-state
analysis of a generic kite moving at constant velocity with the purpose of analyzing the
kite dynamics of the kite itself. Dynamic simulations were introduced to couple the kite
dynamics with the specific geometries and characteristics of the balanced beam kite
power demonstrator design. The dynamic modeling was used to determine the potential
performance, with respect to power output, of the kite power demonstrator in conjunction
with the kite.
In order to prove the concept addressed in this project, some preliminary testing
will be completed in early May 2007. Since the construction of the retraction mechanism
has yet to be completed, a simple counterweight will be utilized to aid the angle of attack
change mechanism in bringing the balanced beam back to the starting position. The use
of the counterweight is merely a simplification of the retraction spring concept that will
be required in future work in order to apply a variable force to the kite in its down stroke
as well as to engage the gear and axle assembly to generate the electrical energy.
103
5.1 Future Recommendations
•
Refine AOA Mechanism
o The design of the preliminary angle of attack change mechanism has been
completed, but testing has yet to be done on the constructed mechanism.
Work may be required to determine the appropriate weight to be
incorporated into the system.
•
Refine Retraction Spring Mechanism
o A preliminary design for the retraction mechanism has been completed,
but due to time constraints the system was not constructed. Though the
design was completed there are several issues that could be addressed.
First of all it will be necessary to find some form of a spring mechanism
that will extend to the required length. Second, if the current design is
taken then the custom ‘chain pulley’ will have to be modeled and milled.
•
Refine Roll Control Mechanism
o The system consisting of springs for the roll control mechanism has been
considered, but little work has been accomplished in the final design of a
system to control the potential roll of the kite in motion. More intensive
work should be done in order to determine if a system will be required and
what components this system could consist of.
•
Finish System Construction – Sprag Clutch
o Due to budget constraints the sprag clutch, which is meant to stop the gear
system from turning when the kite is in its down stroke, was not purchased
and attached to the system. With future budgets a sprag clutch should be
104
purchased and attached to the large sprocket in the gear system. A sprag
clutch would typically be welded to the side of the gear and tightened on
the axle with set screws.
•
Research Different Line Setups
o For this project a standard line setup was adopted to keep the setup as
simple as possible for initial testing. Future testing may incorporate
different line setups to test kite performance. Different setups may be
switching the leading and trailing edge lines or simply trimming the kite
lines to optimize performance.
•
Test Potential Energy Output of Fully Constructed System
o Once the system is fully designed and constructed, a torque meter should
be used to determine the resultant torque on the axle holding the small
sprocket. This test setup will further prove that a kite power demonstrator
has power output capabilities similar to that of a small wind turbine.
•
Extended Testing on Kite Dynamics
o Due to time constraints physical testing of the kite was limited and thus
this portion of the project should be expanded upon in the future. More in
depth analysis of kite dynamics may bring forth significant changes in the
system design when considering the effects of wind gusts and changing
wind direction.
•
Long-term Analysis of Weathering on Kite
o One major concern with a kite-powered system is the effect that weather
may have on the rip-stop nylon material of the kite over extended periods
105
of time. Such concerns that should be considered are UV degradation,
lightning strike sustainability, and ability to withstand extended periods of
rain and sleet.
•
Increased Altitude
o When the concept was first developed in the late 1970’s it was intended
for altitudes approaching 5000 feet. Due to FAA restrictions, a kite in the
U.S. cannot exceed 150 feet in altitude without modifications. We have
yet to test the system and kite together, but aim to test the kite at an
altitude of 80+ feet. In the future it may be beneficial to test the kite at an
altitude approaching the 150’ mark or even higher with the required
modifications.
•
Extended Use of System
o Initial testing of the system will primarily consist of launching the kite
attached to the system and testing it for a short period of time, a number of
minutes. In the future the system should be optimized so as to launch and
fly for an extended period of time, possibly several hours or even a day.
106
References:
1. Goela, J. S. "Wind Power Through Kites." Mechanical Engineering 42 (1979):
42-43.
2. Goela, J. S., R. Vijaykumar, and R. H. Zimmermann. "Performance
Characteristics of a Kite-Powered Pump." Journal of Energy Resources
Technology 108 (1986): 188-193.
3.
Goela, J.S. “Wind Energy Conversion Through Kites.” January 1983.
Indian Institute of Technology Kanpur
4. Goela, J.S., Varma, Sanjeev K. “Effect of Wind Loading on the Design of a Kite
Tether.” Journal of Energy. Vol. 6 No.5 1982
5. Loyd, Miles L. “Crosswind Kite Power.” Journal of Energy. Vol. 4 No.3 MayJune 1980.
6. Nicolaides, John D., Speelman, Ralph J., Menard, George L. C. “A Review of
Para-Foil Applications.” Journal of Aircraft. Vol.7 No.5 Sept.-Oct. 1970.
7. Vijaykumar, R. Performance Characteristics of a Kite-Powered Pump. M. Tech.
Thesis, Department of Mechanical Engineering, IIT Kanpur, Apr. 1984.
8. The Drachen Foundation. 2006. 09 Oct.-Nov. 2006. <http://www.drachen.org>.
9. Lynn, Peter. Peter Lynn Kiteboarding. 2006. Oct-Nov. 2006.
<http://www.peterlynnkiteboarding.com>.
10. “Wind Turbines.” April 2007.
<http://www.rpc.com.au/products/windturbines/wind_faq.html>
11. “Kite Wind Generator.” 2006. April 2007.
<http://www.sequoiaonline.com/blogs/htm/progetto_eng.htm>
12. “Kite –Tugs.” 2003. April 2007.
<http://foxxaero.homestead.com/indsail_028.html>
13. Lang, David. “Electrical Power Generation Using Kites.” Drachen Foundation.
December 2005
107
APPENDIX A: Drachen Foundation
The Drachen Foundation is a non-profit organization dedicated to educating the
public about the history and science of kites. In addition to funding classroom activities
and public events, they also do research into using kites for other means, such as power
generation. The decision matrix used to evaluate our initial concept mechanisms was
based on a similar matrix used in a 2005 paper to compare and contrast the feasibility of
various kite-based power generation concepts.
108
APPENDIX B: Heifer International’s Overlook Farm
Heifer International is a non-profit organization dedicated to educating the public
on poverty and living conditions in under-developed nations. Additionally Heifer seeks
to train people to be self-sufficient and put an end to poverty and hunger around the
world through knowledge and understanding. Overlook Farm is a 270-acre working farm
in Rutland, MA that is open to the public to demonstrate poverty solutions. This site is
the designated test site for the kite power demonstrator, chosen for its proximity to the
WPI campus and its relatively high elevation. Additionally, Overlook Farm resides in a
class 3 wind zone, meaning on average the wind speed is 6.5 to 7 meters per second, or
14.5 to 15.7 mph.
109